Introduction and Hypothesis <p>De novo stress urinary incontinence (SUI) remains a clinically relevant complication after colpocleisis for advanced pelvic organ prolapse (POP), and reliable preoperative risk stratification remains limited. We aimed to develop and internally validate a multimodal prediction model incorporating clinical and pelvic floor ultrasound parameters to estimate the risk of de novo SUI following colpocleisis.</p> Methods <p>This retrospective cohort study included consecutive women undergoing colpocleisis for advanced POP between August 2019 and March 2025. De novo SUI at 1 year was assessed using the International Consultation on Incontinence Questionnaire–Short Form. Candidate predictors included clinical and ultrasound parameters. Independent predictors were identified using multivariable logistic regression. Logistic regression, random forest, support vector machine, and extreme gradient boosting (XGBoost) models were constructed and internally validated using tenfold cross-validation. Model performance was assessed by discrimination, calibration, and decision curve analysis.</p> Results <p>Among 475 eligible women, 237 with preoperative urinary incontinence and 17 with incomplete data or loss to follow-up were excluded. At 1 year, 56 women (25.3%) developed de novo SUI. Age, body mass index, parity, urethral rotation angle, and bladder neck descent were independent predictors. The XGBoost model achieved the highest discrimination (AUC 0.858) and lowest Brier score (0.118). Decision curve analysis indicated favorable net benefit. An online risk calculator was developed.</p> Conclusions <p>A multimodal model incorporating clinical and ultrasound parameters demonstrated good performance for predicting de novo SUI after colpocleisis. The XGBoost model showed the best overall results and may support preoperative risk assessment, pending external validation.</p>